Reviews/Giant Cell Arteritis/Relapse common after tocilizumab discontinuation
Tocilizumab · GCA-TOCI-02 · v2
Borderline: within 4 points of the Established/Likely boundary, so a re-grade could land either side.
When patients with giant cell arteritis stop tocilizumab, a large share of them have their disease come back. Eight separate studies, including follow-up from the main randomised trial [9] and its earlier arm [1], all point the same way: relapse happens often after the drug is stopped. The exact proportion varies a lot between studies, from about a quarter of patients [6][10] to roughly six in ten [3], depending on how long patients were watched and how relapse was defined. The panel treated the claim's word 'common' as a modest bar that the evidence clears even at its lowest reported rate. Most of the studies look back at past records rather than being designed in advance, and they do not define relapse the same way, which is why the panel noted the evidence has limits even while agreeing on the direction.
Across eight included sources the direction is uniform: a substantial fraction of patients relapse after stopping tocilizumab. Most cohorts cluster at 40–62% ([2] >50%, [3] 62%, [4] 47.3% at 18 months, [5] 40% within the first year, [9] 58% not maintaining drug-free remission), while [6] (26% at 6 months) and [10] (25.4%) sit lower, and [1] reports remission persisting in about half after treatment stop. The signal spans North American [2][10], European [3][4][5][6] and randomised-trial [1][9] populations. Even the lowest estimate, 25.4% in [10], supports the modest claim that relapse is common. The contested point: The evidence divides not on direction but on magnitude: reported relapse ranges from 25.4% [10] to 62% [3]. This spread tracks differences in relapse definition (treatment-intensification in [3], CRP/ESR-independent in [4]), withdrawal mode (abrupt 46% vs tapered in [5]), and follow-up length (6 months in [6] versus 18 months in [4]).
All 10 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Giant Cell Arteritis, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
Seven of eight included sources are observational/retrospective with selection and ascertainment confounding; only [9] and [1] derive from randomised allocation.
'Common' is quantitatively imprecise: reported post-discontinuation relapse spans 25.4% [10] to 62% [3] depending on cohort, follow-up duration and relapse definition (CRP/ESR-independent in [4], therapy-escalation in [3]).
Where it remains uncertain. Rates vary more than twofold (25.4% [10] to 62% [3]) because relapse definitions, withdrawal mode [5], and follow-up windows [4][6] differ, and seven of eight sources are observational.
Aleatory
Even within a cohort, who relapses and when varies widely; [1] records relapses from 2 to 14 months and [10] a median of 283.5 days, reflecting irreducible patient-to-patient variation.
Epistemic
Seven of eight sources are retrospective/observational [2][3][4][5][6][7-excluded][8-excluded][10], so selection and ascertainment bias limit how firmly the rate is known; only [1][9] are randomised-derived.
Inferential
'After discontinuation' pools abrupt stops (46% in [5]) with tapered withdrawal (26% in [6], 11–16% while still on reduced TCZ in [5]), so the aggregate obscures exposure-dependent risk.
The load-bearing papers from a corpus of 10, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
About half relapse after a 52-week TCZ course is stopped.
“The data show that a 52-week treatment with tocilizumab induces a lasting remission that persists in half of the patients after treatment stop.”
Panel auditInternal validity: medium · scope of inference: GCA patients who achieved lasting remission on a 52-week tocilizumab course, followed after stopping.Key limitations: very small sample · single-centre/single-trial arm · no comparator
27 of 52 patients relapsed after stopping TCZ, 58% within 12 months.
“Fifty-two patients stopped TCZ after a median of 16.8 months; 27 relapsed after discontinuation (median: 8.4 months; 58% relapsed within 12 months).”
Panel auditInternal validity: medium · scope of inference: North American GCA patients treated with tocilizumab in routine care who later stopped it.Key limitations: retrospective · single enterprise · discontinuation reasons heterogeneous
62% relapse rate after discontinuation in a French multicentre cohort.
“Among 29 patients undergoing TCZ discontinuation, 18 (62%) experienced relapses.”
Panel auditInternal validity: medium · scope of inference: Elderly, heavily pre-treated refractory GCA patients at three French centres who discontinued tocilizumab.Key limitations: small sample · retrospective · refractory/selected population
3 of 10 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Uniform direction across eight independent cohorts and an RCT extension clears the low bar of 'common', even at the lowest observed relapse rate of 25.4%.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Randomised evidence on discontinuation. Only [9] and the 20-patient [1] derive from randomised allocation, so the headline rate rests largely on retrospective data. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A randomised continuation-versus-discontinuation trial after a fixed TCZ course, with blinded relapse adjudication.
Open question: What is the pooled post-discontinuation relapse rate under a single harmonised definition and fixed follow-up? Why it is still open: Each cohort uses its own relapse definition and follow-up window ([3] therapy escalation, [4] CRP/ESR-independent, [6] 6 months, [4] 18 months), so no comparable pooled estimate exists.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19